8 Market Approval for Drug; and Medical Devices

Market Approval for Drug;
and Medical Devices
Introduction and Background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
FDA’s Prescription Drug Market Approval Process . . . . . . . . . . . . . . . . . . . . . . . . 86
FDA’s Medical Device Evaluation Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Safety and Efficacy Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Current Economic Considerations in FDA’s Drug and Medical Device Approval
Processes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Potential Application of CEA/CBA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Implications of CEA/CBAin Market Approval: General Findings . . . . . . . . . . . . . 94
Potential Positive Effects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Potential Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Market Approval for Drugs
and Medical Devices
The Federal Government has been authorized
to regulate various aspects of drugs in the
United States since President Theodore Roosevelt signed into law the Pure Food and Drug Act
of 1906 (113). That law was in part intended to
help prevent adulteration and misbranding of
drug products, In 1912, Congress enacted the
“Sherley amendment,” which prohibited companies from making false and fraudulent curative or therapeutic claims on the labels of their
products. In 1927, Congress established the
Food and Drug Administration (FDA).
The 1938 Food, Drug, and Cosmetic Act,
passed by Congress in response to a tragic event
in which over 100 people died from ingesting a
lethal drug product, ’ increased the Federal Government’s regulatory control over the marketing
of drugs, devices, foods, and cosmetics. In addition to establishing several labeling requirements for drugs, the 1938 Act prohibited interstate commercial shipment of new drugs until
they had been adequately evaluated by the Fed-
eral Government to show that they were safe
under the conditions of use listed on their label.
It also authorized FDA to remove from the market any drug it could prove to be unsafe. The
Durham-Humphrey Amendment of 1951 defined criteria and categories based on levels of
drug safety for restricting a drug to legend (i.e.,
prescription only) status.
Again stimulated by a disaster, this one involving fetal abnormalities caused by the drug
thalidomide, z Congress enacted the Drug
Amendments of 1962 (Kefauver-Harris amendments). These amendments required drug manufacturers to provide “substantial evidence”
that their products were efficacious as well as
safe. The amendments also required drug manufacturers to report promptly to FDA information concerning the safety and efficacy of their
marketed products and strengthened FDA’s authority to remove unsafe or ineffective drugs
from the market. In addition, the 1962 amendments authorized the notice of claimed investigation for a new drug (lND) process—a legal
mechanism used by FDA to regulate human investigations of drugs.
The Federal Government was given some authority to regulate medical devices in the
original 1938 Food, Drug, and Cosmetic Act.
Congress substantially increased this Federal
regulatory authority, however, by passing the
Medical Device Amendments of 1976 (116). The
1976 amendments authorize FDA to require device manufacturers to demonstrate acceptable
levels of safety and effectiveness for their prod-
ucts. They establish three categories of products, and each category has a different level of
Federal control. Medical devices that are implanted, used in supporting or sustaining human
life, of substantial importance in preventing impairment of human health, or that pose a potential unreasonable risk of illness or injury, for example, usually are subjected to premarketing
evaluation of safety and efficacy. The 1976
amendments permit FDA to ban or remove
from interstate commerce unsafe or ineffective
devices. They also allow the Agency to restrict
the use of a device to persons with specific training or experience or to those in specified facilities if the device’s safety and effectiveness cannot be assured in general use.
Since 1962, FDA has used its statutory and
regulatory authority to establish an extensive
system for evaluating virtually every new prescription drug prior to its release into the general medical marketplace. By law, FDA must base
its evaluation of prescription drugs on two criteria: safety and efficacy.3 The procedures used
by FDA in this premarket approval process, described in detail elsewhere (119,144), are briefly
summarized below. To initiate human investigations involving agents legally defined as “new
drugs, ” the drug “sponsor” (e.g., a manufacturer or an independent investigator of a new
drug) must obtain FDA’s approval. To obtain
such approval, the sponsor submits to FDA’s
Bureau of Drugs (BOD) an IND, describing the
qualifications of the investigators and the design
of the planned trials. In the IND review process,
BOD also examines data regarding the pharmacology and toxicology of the applicant drug
collected in animal studies and in human studies
that were not subject to FDA approval (e.g.,
those conducted in foreign countries). If the
sponsor’s IND is approved, the sponsor may
proceed with clinical testing (i.e., testing in
human subjects) in the United States, After completing clinical testing under IND, the sponsor
files with BOD a new drug application (NDA)
that describes in detail the results of IND clinical
trials. The applicant drug has usually been
tested in 500 to 3,000 human test subjects (306).
By filing an NDA, the sponsor is requesting
FDA’s permission to market the new drug in interstate commerce.
According to Dr. Marion J. Finkel, Associate
Director for New Drug Evaluation, BOD, FDA,
the review process for INDs and NDAs proceeds
as follows (206):
‘Food, Drug, and Cosmetic Act S505, 21 U.S. C. S 355 (1976)
part 314 (1979). The Act uses the term “effectiveness” instead of
“efficacy. ”
Each IND and NDA application is reviewed
by a team of FDA scientists: a physician, a pharmacologist, a chemist, a pharmacokineticist,
usually a biometrician, and when applicable, a
microbiologist. Important NDA’s are then presented for consideration to advisory committees, of which the Bureau of Drugs has 13, consisting of extramural [mostly nongovernmental]
experts in, principally, the subspecialities of
medicine, clinical pharmacology, and biometrics. The committees recommend to the FDA
whether or not an NDA should be approved for
marketing and, if so, under what labeling, and
whether the sponsor should be requested to perform additional studies in the postmarketing phase. When recommendations are made
against approval, the committees provide advice
on new studies to be done by the sponsors to explore the drug’s safety and effectiveness further.
The use of advisory committees is the FDA’s primary method for broadening the decisionmaking process.
During the NDA review process, FDA reviewers analyze the sponsor’s summaries of
data—and when needed the actual raw data—
generated from clinical investigations. A major
task for BOD during this review process is to ensure that clinical experimental data support the
sponsor’s claims for the drug’s safety and efficacy that appear in the drug’s labeling. The
NDA review process usually entails deliberations between FDA and a drug’s sponsor regarding claims for safety and efficacy. If FDA concludes that an NDA is deficient, it usually will
require the sponsor to generate new data, modify its NDA with additional information, and resubmit the application to the Agency. Once a
sponsor obtains an NDA approval from FDA, it
is authorized to market the drug in interstate
commerce for only the specific indications (uses)
that have been authorized by FDA.
The Medical Device Amendments of 1976
(Public Law 94-295) require FDA’s Bureau of
Medical Devices (BMD) to classify each medical
device—on the basis of the level of regulation
necessary to ensure safety and efficacy—into
one of three regulatory classes. Products placed
in the class I category —those requiring the least
controls to ensure their safety and efficacy—are
subject only to general controls, including premarket notification, registration of the manufacturer, prohibition of product misbranding
or adulteration, adherence to FDA-promulgated
good manufacturing practices, and compliance
with various recordkeeping requirements. Class
II products—those for which general controls
are deemed inadequate to ensure their safety
and efficacy—must comply with performance
standards either established or recognized by
BMD when existing information permits development of such standards. General controls also
apply to class II devices unless superseded by a
standard. Class III devices—those for which
neither general controls nor performance standards alone are sufficient to ensure their safety
and efficacy—are subject to premarketing approval. (BMD can also require premarketing approval of devices for which insufficient information is available for the development of applicable performance standards. ) Class III devices are subject to general controls.
The process BMD uses to evaluate the safety
and efficacy of new class 111 medical devices is
similar to that used by BOD in its NDA reviews.
BMD, however, is required by law to use its advisory panels during a product’s review,
whereas BOD’S use of advisory panels is discretionary. BMD is also authorized to use another
premarketing approval process, the product development protocol (PDP). PDP was designed
to encourage the development of, and to streamline the approval process for, innovative medical devices. During a PDP process, investigation
of a device and the development of information
necessary for its premarket approval are merged
into one regulatory mechanism. BMD works
closely with the manufacturer while information to support the device’s safety and effectiveness is being developed. The 1976 Medical
Device Amendments also include an investigational device exemption (IDE) provision. IDE
allows FDA (technically, the Secretary of
Health and Human Services) to exempt a new
device from other provisions of the amendments
to permit controlled testing of new devices prior
to commercial marketing. The IDE process is
similar to the IND process that BOD uses.4
45 FR 3732.
The Food, Drug, and Cosmetic Act mandates
FDA to require a drug sponsor to: 1) collect, by
all methods reasonably applicable, data that
demonstrate a drug’s safety, and 2) generate
“substantial evidence” from controlled trials to
show that the drug is efficacious for use under
the conditions set forth in the proposed labeling
(144). Although the Act provides no definitions
and little guidance for the meanings of safe and
effective, 5 it does describe “substantial evidence” of effectiveness as follows (144):
The term “substantial evidence” means evidence consisting of adequate and well-controlled
investigations, including clinical investigations,
by experts qualified by scientific training and experience to evaluate the effectiveness of the drug
involved, on the basis of which it could fairly
and responsibly be concluded by such experts
that the drug will have the effect it purports or is
represented to have under the conditions of use
prescribed, recommended, or suggested in the
labeling or proposed labeling thereof.
Because no drug is absolutely safe or always
effective under all conditions of use, FDA
weighs a drug’s benefits in comparison to its
risks. In FDA’s evaluation process, an applicant
drug’s safety and efficacy are compared to the
safety and efficacy of approved products or
medical procedures used to achieve similar therapeutic objectives. Comparisons are also made
with the effects of inert substances (placebos).
According to Dr. Marion J. Finkel of FDA (204):
A drug cannot be considered safe if it is less effective and has more side effects than other
drugs labeled for the same indication. It might
be safe for certain subpopulations or certain restricted indications, however.
According to Dr. Robert Temple of FDA (594):
We’re not supposed to refuse approval because of lesser effectiveness, but we can on
grounds of lesser safety which reduces the benefit/risk ratio. Relative effectiveness can be noted
in labeling.
FDA also uses such comparisons when selecting
NDAs for its priority review or “fast track” review process.
FDA has published 25 clinical guidelines that
define the types of studies it deems appropriate
to use to establish safety and efficacy of a drug.
The performance criteria used to evaluate an individual drug’s safety and efficacy, however,
can vary among different pharmacological
classes. This variance results from the imprecise
nature of pharmacologic intervention in disease.
The benefits and risks of a drug can vary from
disease to disease, from population to population, and from clinical situation to clinical situation. Drug-induced hair loss, for example, is
often viewed as an acceptable risk of taking a
drug that is an effective treatment for cancer,
especially when no other effective therapy exists. The same side effect, however, would probably not be acceptable for a drug that reduces
the severity of a minor, self-limiting condition.
The complete risk-benefit analysis of a drug,
therefore, is not based solely on statistical
evaluation of data on safety and efficacy generated from clinical trials but also on “the context
of the disease for which [the drug] is intended,
the availability of other therapeutic modalities,
including other forms of pharmacologic therapy, and public health implications of its availability” (144).
The task of evaluating the safety and efficacy
of drugs, as well as FDA’s ability to perform this
task, has been subjected to extensive analysis
and debate (119,144,242,613).
FDA does not directly use economic criteria
in its approval processes for drugs and medical
devices. The Agency’s primary statutory authority, the Food, Drug, and Cosmetic Act, neither authorizes nor prohibits the use of economic criteria in FDA’s evaluation of applicant
drugs and devices. The legality of using cost effectiveness to help evaluate new drugs and devices has not been tested.
Although FDA does not formally assess the
potential economic impact a drug or medical device might have on the allocation of health care
resources, some FDA actions may be based indirectly on —or taken in spite of—economic considerations. Further, some FDA actions certainly have economic impacts. Several examples are
cited below.
Ch 8—Market Approval for Drugs arid Medical Devices
Classifying the Potential Therapeutic Importance of New Drugs: FDA sets priorities for its
review of new drugs according to the potential
therapeutic usefulness of each applicant drug.
This priority-setting process supposedly begins
within 6 months after a sponsor submits an IND
to FDA (206). On the basis of preliminary in-
formation about a drug’s pharmacological effects and on data from clinical trials, a drug is
assigned a number and a letter derived from the
following classification schemes (221):
Numerical Classification
New molecular entity not previously
marketed in the United States.
New salt, ester, or derivative of an active
moiety marketed in the United States.
New formulation of a compound marketed in the United States.
New combination of two or more components not previously marketed together in the United States.
Product duplicates a drug marketed by
another firm in the United States.
Product already marketed by some firm
in the United States; approval being
sought for new indication of use.
Letter Classification
A = major therapeutic advance over other
currently available drugs, etc.
B = modest (or moderate) therapeutic advance over other currently available
drugs, etc.
c= little or no appreciable therapeutic advance over other currently available
drugs, etc.
The purpose of using a drug classification
scheme-of this type is to expedite the review of
data for drugs that represent important new
therapeutic entities. FDA seeks to review NDAs
submitted for drugs it assigned a 1A or IB rating
as expeditiously as possible; such reviews receive high priority by BOD (206).
Even though the criteria used to classify a new
drug are based on scientific data relating to a
drug’s clinical safety and efficacy, there are indirect economic consequences of FDA’s selection process. The expeditious review and resultant early marketing of a new drug that repre-
sents a major therapeutic breakthrough in the
treatment of a heretofore uncontrollable disease
could help reduce the use of existing treatment
measures, such as hospitalization and surgery.
The recently introduced drug cimetidine, for example, appears to provide safe, effective, and
relatively inexpensive (compared to hospitalization or surgery) treatment for duodenal ulcers;
using cimetidine, some ulcer patients may avoid
hospitalization (203,632). FDA rated cimetidine
as a 1A drug. If indeed the use of cimetidine
reduces ulcer patients’ hospitalizations and
surgery, then an expeditious FDA review and
approval of the drug could help reduce medical
expenditures for the treatment of duodenal
ulcers. A complete analysis would include calculations of the potential economic impacts of
delaying the review of one or more other INDs
or NDAs, in order to expedite cimetidine’s application review.
According to FDA’s Dr. Marion J. Finkel
NDAs for 1A and IB drugs are full NDAs,
containing all of the safety and efficacy data required for any NDA. Expeditious review of
these merely means that the NDAs do not wait
their turn in the pipeline but are reviewed before
NDAs with lesser classifications . . .
Rarely, FDA will accept an NDA for a 1A
drug without as much long term human (or animal) safety data required for NDA approval.
In at least some cases, however, the expeditious review of an important new drug might result in an incomplete assessment of the drug’s
safety, and an unexpectedly significant level of
adverse reactions to the drug might occur. To
help prevent such an occurrence, FDA can—and
sometimes does—ask a sponsor to conduct postmarketing surveillance as a condition of approval for a new drug (131).
Use of Abbreviated New Drug Applications
(ANDA) and “Paper” NDAs: Two examples of
the indirect use of economic considerations in
FDA’s drug approval processes are the ANDA
and the so-called “paper” NDA. The ANDA
process enables a drug manufacturer to obtain
from FDA marketing approval for a “generic”
drug product that is purported to be an identical
version of an already approved product, usually
after the originator’s patent has expired. Because of FDA’s earlier interpretation of the
Food, Drug, and Cosmetic Act, ANDAs are currently only used for products originally approved by FDA between 1938 and 1962. FDA is
preparing a policy, however, that will permit
the use of ANDAs for products approved after
1962. A manufacturer can obtain an ANDA essentially by demonstrating that it complies with
FDA’s current good manufacturing practices
and labeling requirements and that it can make
a product that is at least chemically equivalent
and supposedly bioequivalent to the originators. Although evidence of chemical equivalence is always required, the FDA Commissioner can forgo requiring evidence of bioavailability in the ANDA process for a given drug.
“Paper” NDAs are designed to permit a drug
manufacturer to meet the requirements for submitting evidence of the safety and efficacy of its
post-1962 product by citing existing data from
published clinical trials involving a chemically
equivalent, previously approved product. The
legality of “paper” NDAs was challenged in
court by the Pharmaceutical Manufacturers Association (PMA) and a few major drug manufacturers (473). The court dismissed the suit on
the grounds that the plantiffs had not exhausted
their administrative remedies within FDA, so
PMA has filed a petition that: 1) questions
FDA’s authority to implement a “paper” NDA
policy, and 2) requests that if such a policy is
implemented by FDA, it be done through the official notice and comment rulemaking procedure.
The primary purpose of both the ANDA and
“paper” NDA modifications in the new drug approval process is to facilitate the marketing approval of drug products identical to those that
FDA has previously found to be safe and efficacious on the basis of data supplied in a product’s
original NDA. These mechanisms are designed
to prevent unnecessary work burdens on FDA’s
resources. The use of ANDAs and “paper”
NDAs is supposed to increase the availability of
FDA staff to review NDAs submitted for innovative drugs, e.g., those categorized by FDA as
1A. The use of these mechanisms also eliminates
the need for sponsors to conduct duplicative
clinical trials. Thus, manufacturers’ expenses
associated with entering some existing drug
product markets can be reduced.
Removal of Diethylstilbestrol (DES) From
Animal Feed: The estrogenic compound DES
was used by poultry and other livestock raisers
for many years to increase the muscle and fat
content of their animals. FDA first approved
such use of DES in 1954. 6 After that, DES was
used in animal feed to reduce the total feed
intake necessary to achieve maximum animal
weight gain. Furthermore, some studies showed
that the use of DES in animal feed helped reduce
the cost of raising poultry and other livestock.
Theoretically, therefore, the use of DES may
have helped contain poultry and meat prices for
Ingestion of DES by humans, however, was
eventually correlated with an above-normal risk
of developing certain types of cancer, such as
adenocarcinoma of the vagina or the cervix.7
Because of DES’s cancer-producing potential,
FDA attempted several times to remove this
substance from animal feed.
The authority for its efforts was the Delaney
clause,8 a proviso added to the 1958 food additive amendment to the Food, Drug, and Cosmetic Act:
. . . no additive shall be deemed to be safe if
found to induce cancer when ingested by man or
animal . . .
This clause was reiterated in the 1968 animal
drug amendments to the same Act.’ Under the
animal drug amendments, approvals for the use
of new animal drugs, which are used in the livestock industry for the treatment and prevention
of disease and as growth promoters, are to be
granted only after a two-part evaluation by
FDA, First, FDA must determine that the drug is
safe and effective for use in animals. Second, it
must assess the safety of potential residues that
might occur in food derived from treated animals. 10 —.
“44 FR 54853.
?44 FR 54852.
‘Sec. 409(c)(3)(A), Food, Drug, and Cosmetic Act, 21 U. S.C.
‘Sec. 512, 21 U,S.C. 360b.
los ec, 512, 23 U. S.C. 360 b(d)(l ).
After evaluating and finding DES to be a carcinogenic substance, in 1972, FDA attempted to
ban its use in animal feed. That effort termi-
nated in a litany of legal proceedings with the
livestock industry and animal feed manufacturers. ” On January 12, 1976, therefore, FDA once
again initiated a legal process to remove DES
from animal feed. This time, the Agency also
issued an inflation impact statement regarding
its proposed ban of the use of DES. I* The statement concluded:
There are no satisfactory alternatives to
the Agency’s proposed action which are
consistent with the legal constraints imposed by the Federal Food, Drug, and
Cosmetic Act as amended.
Operating expenses to feedlot producers
of cattle for feed and other items are estimated to decrease by $156 million during
the first year following a DES ban. These
increased costs are expected to fall substantially over the longer term as substitutes to DES become available in greater
Retail prices of beef are estimated to rise
by about 2 cents per pound, meaning the
per capita cost of beef to consumers at
current levels of consumption would increase from $2 to $3 annually. The aggregate consumer impact is estimated at $.503
A ban on DES would not cause major inflation impacts, as defined by the HEW/
OMB criteria, in the areas of competition,
productivity, supplies of materials, or use
of energy.
The benefits from implementing the pro-
On June 29, 1979, then FDA Commissioner
Donald Kennedy ordered the withdrawal of approval of the new animal drug application for
DES. *3 In taking this action to ban the use of
DES, Commissioner Kennedy stated:
FDA is not authorized, under the Food, Drug
and Cosmetic Act, in considering the question of
whether a new animal drug has been shown to
be safe for use, to weigh the “socio-economic”
benefits that that drug provides against a health
risk to the ultimate human consumers of treated
animals. Even were I to attempt to weigh the
benefits of DES against its risks, this record
would not provide sufficient information to
compute the risk associated with DES or to determine whether, and to what extent, use of DES
provides any health benefit or even any economic benefit to society.
This case study illustrates one situation in
which FDA interpreted the Delaney clause of the
Food, Drug, and Cosmetic Act (its primary statutory authority) to mean that in its decisions
concerning the removal of carcinogenic substances from the market, it cannot consider the
results of economic analyses.
Evaluation of an X-Ray Equipment Performance Standard Established by the Bureau of Radiological Health (BRH): This case illustrates
FDA’s use of cost-benefit analysis to evaluate
one of its regulations.
In 1972, FDA’s BRH promulgated regulations
containing a performance standard for medical
diagnostic X-ray systems and their major components. 14 These regulations, among other
things, required X-ray equipment manufacturers to certify that all specified components manufactured after August 1, 1979, comply with the
new standard. Another provision stated that
after August 1, 1974, no uncertified components
could be assembled or reassembled into an X ray system. 15 The purpose of these regulations
was to help ensure the radiation safety of diagnostic X-ray systems. ‘b
In 1974, BRH estimated the costs and benefits
associated with these regulations. The Bureau
estimated, for example, the expected numbers
and costs of X-ray systems which would be marketed under the new performance standards.
Upon reviewing 3 years of experience with the
regulations for diagnostic X-ray systems, BRH
realized that their earlier estimates of certain
costs and benefits were incorrect. In 1978, using
data not previously available, BRH conducted a
second analysis of costs and benefits derived
from the 1972 regulations. On the basis of this
second analysis, BRH concluded: 17
1. The provisions of section 1000.16 that are
effective after August 1, 1979, will affect
only a small fraction of the uncertified Xray systems currently in use.
2. The total impact of this regulation, in
terms of increased cost for the X-ray
equipment or interruption of health care
delivery, may be significant.
3. For those uncertified systems that would
be affected, by virtue of their sale and relocation, section 1000.16 is not likely to
be a cost-effective approach to improve
the radiation safety performance of X-ray
Based on these conclusions, BRH, through the
rulemaking process, modified its earlier regulations by, among other things, permitting the installation of uncertified components into existing systems whose components are all uncertified and permitting the continued reassembly of
uncertified equipment. It also clarified certain
aspects of its performance standard for X-ray
systems. 18
‘f144 FR 49667.
Incorporating Costs-Savings Studies Into
Vaccine Approval Decisions: In its recently
published report on viral and rickettsial vaccines, FDA’s Panel of Review of Viral and Rickettsial Vaccines identified economics as a major
consideration in the evaluation of vaccines .19
The panel noted that because adequate data are
often lacking, economic considerations are
often ignored in such evaluations.
The panel stated:
These data would contrast the cost of a vaccine and its administration plus the costs (medical care, rehabilitation, impairment of ability to
earn income) of vaccine-related disease with
costs of a similar nature that would have accrued from cases of the natural disease.
The panel further stated:
For an “acceptably safe” and effective vaccine
against a serious disease, , . . the ratio should
be highly favorable. However, if the preventable
disease occurs chiefly in young children and is
infrequently associated with permanent sequelae, a different answer might result. The question then might become “how much cost can be
justified to prevent one crippling or lethal case of
disease?” This clearly requires societal rather
than scientific judgment.
This panel has strictly an advisory capacity
within FDA, and its reports do not reflect official FDA policy, The panel’s discussion of economics in the vaccine evaluation process, however, does illustrate some concern among FDA’s
advisors for evaluating the potential economic
impacts of vaccines.
’45 FR 25665.
In order to facilitate the discussion presented
in the next section of this chapter on the implications of including cost-effectiveness criteria in
the market approval process, OTA has developed the hypothetical model of a cost-effectiveness analysis (CEA) outlined below. It is very
important to keep in mind that the use of the
model for illustrative purposes does not negate
or diminish any of the weaknesses or possibilities for misuse involved in formal cost-effec-
tiveness analyses/cost-benefit analyses (CEA/
CBAS) that seek a single, quantitative bottomline, such as a cost-effectiveness ratio. Such a
ratio is used in the model because 1 ) it points out
some of the data and other problems related to
use of bottom-lines, and 2) it is possible, given
the quantitative nature of FDA’s regulatory
process, that use of a CEA ratio might be seriously considered by the agency if CEA were
added to its mandate.
Ch. 8—Market Approual for Drugs and Medical Devices ● 93
OTA’s model is simplistic by intent and does
not address the numerous assumptions and variations that are possible. Several caveats and
assumptions should accompany an analysis
such as the one presented. Potential methodological problems associated with the use of
CEA in general are explained in detail in a
background paper of this assessment.zo Additional problems with its use in market approval
processes are explained below. One problem,
for example, is that data for some important
variables would be difficult, if not impossible,
to obtain.
In assessing a drug’s cost effectiveness, FDA
might first assess the agent’s efficacy and safety
and quantify its effects into measurable units of
“net health effect. ” Let us assume, for example,
that an NDA is submitted for applicant drug
“D,” which is used to treat high blood pressure
(hypertension). Let us also assume that in premarketing clinical trials, drug “D” consistently
lowered by 10 percent blood pressure in 50 percent of tested hypertensive patients, and that
such a drop in pressure could be correlated with
a 5-percent reduction in morbidity and mortality (e.g., heart attacks, strokes, and kidney disease). Suppose drug “D” also produced undesirable side effects that in premarketing clinical
trials accounted for a 2-percent increase in days
of disability in the tested population. Let us also
assume that the tested population accurately
represents the general hypertensive population.
To calculate the “net health effect” of drug
“D,” all such positive effects (e.g., 5-percent
reduction in mortality) and negative effects
(e.g., 2-percent increase in disability caused by
side effects) would have to be converted into a
uniform and measurable unit of health .21 The
“net health effect” of drug “D” could then be
calculated for each indication listed on the proposed drug labeling (e.g., treatment of moderate
hypertension in ambulant patients). Specific
conditions of use (e. g., in conjunction with
other drugs) and peculiar effects in special pop-
ulations (e. g., the elderly) could be accounted
for in the calculation of net health effects.
Once the drug’s effects were converted into
measurable units of “net health effects, ” the “net
cost” of achieving a desirable level of health effect (e.g., a 5-percent reduction in mortality and
morbidity) would have to be calculated. A “net
cost” could include such items as the cost of purchasing the drug, the cost of treating drug-induced side effects, and perhaps the cost of treating other illnesses in the persons whose lives are
saved by the use of drug “D. ” Subtracted from
such costs could be savings resulting from any
reduced costs of hypertension treatment (e. g.,
lower use of previously approved drugs or decline in hospitalizations and in physician office
visits) resulting from the use of applicant drug
At the conclusion of this phase of the analysis, one could construct a ratio of net cost in
dollars to one unit of “net health effect”
achieved through the use of drug “D” in the
treatment of hypertension, e.g.:
Cost-effectiveness ratio =
Net cost
Unit of net health effect
In the next phase of the analysis, the net cost
(in dollars) of achieving a desired net health effect through the use of drug “D” would be compared with the net cost of achieving the same net
health effect by using an existing approved
treatment modaIity (e.g., another drug, surgery, or biofeedback) to lower blood pressure.
Such a comparison of costs would require that
cost-effectiveness ratios, i.e., net cost (in dol1ars)/unit of net health effect, be derived for
each hypertension treatment modality. The following hypothetical ratios for example, might
be derived:
Hypertension treatment
Applicant drug “D”. . . .
Approved drug “A” . . .
Surgical procedure “X” .
Biofeedback . . . . . . . . .
Net cost/uniform unit
of net health ef}ect
Let us assume that no other treatment of hypertension was available at the time of the analysis.
The final phase of such an analysis would be
to establish criteria for judging the cost-effec-
tiveness ratios for each treatment modality and
for using such ratios to help determine if applicant drug “D” should be approved. In the example above, if a criterion for market approval
were that the applicant drug had to produce one
unit of “net health effect” at a cost lower than
the cost of using any other approved antihypertensive drug, then drug “D” might not be approved. If such a criterion included comparison
with other approved forms of hypertension
treatment, then drug “D” would presumably be
approved if compared with surgery, although it
might not be approved if compared with biofeedback. When one considers in this hypothetical evaluation process the need for a variety of
available treatment modalities to meet individualized patient situations, then the relatively
small differences in cost effectiveness between
applicant drug D, biofeedback, and approved
drug A become much less important.
Cost effectiveness might be used in formal review processes of drugs that have been marketed for a number of years, instead of those for
new drugs. FDA conducts at least three such re-
view processes, and to date, the Agency has not
used cost effectiveness in any of them. First, in
its drug efficacy study implementation project,
FDA is conducting a one-time review of the efficacy of drug products approved between 1938
and 1962 (131). The Agency categorizes products according to their documented clinical efficacy, has removed from the market some
products for which efficacy documentation is
lacking (e.g., selected fixed-dosage antibiotic
combination products), and is attempting to remove other such products. Second, in its Overthe-Counter (OTC) Drug Review Program,
FDA is conducting a one-time review of the safety and efficacy of several hundred ingredients
used in OTC or nonprescription drug products.
Third, FDA (technically, the Secretary of Health
and Human Services) has authority to remove
from the market a drug that represents an “imminent hazard” to the public’s health. When this
authority was exercised to remove the antidiabetic drug phenformin from the market, FDA
extensively reviewed the drug’s safety and to
some extent its efficacy.
As described in chapter 2, CEA is an analytical device that can be used to help decisionmakers allocate resources, usually in the public
sector. The primary purpose of FDA’s drug and
medical device market approval processes is to
protect the public from unsafe and ineffective
products. FDA’s marketing approval processes
do, however, indirectly influence the allocation
of public resources in at least two ways. First,
the types of regulations and procedures established by FDA affect the allocation of its operating budget. Second, many—if not most—of the
drugs and medical devices approved by FDA
eventually are used in health care services that
are paid for through publicly financed programs, such as medicare and medicaid.
The use of cost effectiveness as a criterion in
FDA’s market approval processes for drugs and
medical devices would require a substantial shift
in the Federal Government’s approach to regulating the medical care marketplace. At present,
no Government effort is designed explicitly to
reduce medical expenditures by directly slowing
down or stopping the market introduction of
medical goods and services. The Federal Government does not extensively evaluate the cost
effectiveness of drugs or medical devices in any
of its efforts to contain health care costs. Conceivably, such evaluation could take place in the
National Center for Health Care Technology,
which advises the Health Care Financing Administration (HCFA) on reimbursement policies
for selected medical technologies. Results of
CEAS involving drugs and medical devices
could be incorporated into HCFA’S policies for
reimbursing the use of selected drugs and medical technologies under medicare, medicaid, and
any other federally operated health insurance
Ch. 8—Market Approval for Drugs and Medical Devices
Potential Positive Effects
Potential Problems
There are at least two hypothetical positive
effects of incorporating CEA into FDA’s market
approval process for drugs and devices.
Among the consumer advocates, FDA employees, and representatives of the pharmaceutical industry surveyed in this assessment, there
appeared to be widespread agreement that cost
effectiveness is not an appropriate criterion to
use in the drug and medical device approval
process. Using information and ideas obtained
from several individuals, OTA developed the
following analysis of potential problems.
First, manufacturers could formally incorporate cost-effectiveness criteria, based on societal
values, into their research, development, and
marketing strategies. Manufacturers most likely
do use some form of CEA/CBA to allocate their
R&D expenditures. The primary criteria used in
such allocations, however, may emphasize such
items as total sales, market portions, and return-on-investment. If FDA used criteria such as
reduced treatment costs, improved levels of
health, and improved efficiency in disease prevention or treatment, then more manufacturers
might develop new products and seek new markets where existing treatment or prevention
measures are ineffective or inefficient. Some industry representatives claim that the leading
drug research firms currently include such public health criteria in their research priorities. The
existence of orphan drugs (i.e., existing agents
for which there is a small, demonstrated clinical
need but no manufacturer), however, indicates
that at least for certain products, profitability
takes precedence over societal need. Conceivably, such use of CEA could help the Government encourage participation by drug and medical device manufacturers in public efforts to develop more cost-effective medical technologies.
Second, if FDA were able to accurately assess
the cost effectiveness of an applicant medical
device or drug, the Agency might help reduce
expenditures for inefficient products by keeping
them off the market entirely. By evaluating the
cost effectiveness of medical devices and drugs
in the market approval process, the Government would be assessing a product very early in
its diffusion process. Perhaps this early evaluation process would lead to better direction—and
perhaps an expansion—of experimentation with
new drugs and devices. At present, substantial
non-FDA-approved clinical experimentation
with newly approved drugs takes place in uncontrolled situations. Such early evaluation,
however, would have no effect on the inefficient
use of drugs and medical devices found to be
cost effective in clinical trials.
First, the market approval process may be too
early in the life of a drug or medical device to
evaluate its cost effectiveness; information regarding a product’s safety and efficacy are usually available for only one or two indications of
use at the time a sponsor submits an initial NDA
for a new drug or an application for approval of
a new device. Data regarding a drug’s safety and
efficacy in medical conditions not listed in the
product’s official labeling are not often generated until a product has been marketed for at
least a few years. If a new drug or medical device were not approved for marketing because
its cost effectiveness did not compare favorably
with already approved products, then the new
product might never be fully evaluated, particularly in the treatment of medical problems other
than the one(s) studied initially. The total benefits, risks, and cost effectiveness of drugs that
are used in the treatment of more than one medical problem often may not be known for several
years after the product’s initial development.
Examples of such products include propranolol
used in the treatment of angina, migraine headaches, and hypertension; selected antibiotics
used in the treatment of acne; amantadine used
in the treatment of parkinsonism; and phenytoin used in the treatment of certain cardiac arrhythmias. In addition, rare or delayed-onset
adverse drug reactions, for example, drug-induced cancers, would not likely be included in
premarketing CEA.
James B. Russo, SmithKline Corp., offers an
industrial concern about the use of CBA in the
market approval process for drugs (529):
Our primary concern is over the fact that costbenefit analysis in the drug area is a new discipline, and one which simply must not be widely
96 ● The Implications of Cost-Effectiuertess Analysis of Medical Technology
applied, particularly in the area of new drug approval, until we have real confidence in its predictive reliability . . . I don’t really know how
to state that point as strongly as I’d like to.
Think back to probenecid. The drug was developed to slow the excretion of penicillin from the
kidney, because in those days penicillin production was inadequate. By the time probenecid
was shown to be safe and effective for that purpose, penicillin was coming out of the industry’s
ears. Had that NDA been looked at on the basis
of the product’s cost against its possible benefits
in prolonging blood levels of penicillin, work on
the drug would have been stopped in the early
‘50s. Of course, once it was found that it speeded
the excretion of uric acid, an entirely new and a
relatively important means of relieving gout became a possibility, It isn’t simply a matter of
NDA approval. If we knew we would have to
pass that kind of test at the NDA stage, I fear
that a lot of projects would be cancelled long before we had practical information on the drug’s
full potential.
In recent years, however, FDA has approved
very few NDAs to add to a previously approved
product’s official labeling a new major indication that required data from new clinical trials.
Between January 1, 1974, and September 30,
1979, for example, FDA approved 4 such NDAs
out of a total of 484 total NDA approvals for all
reasons (221). There are, however, 362 active
commercial INDs for products being used for indications not currently listed in the products’ official labeling. In addition, for perhaps hundreds of approved products, FDA has permitted
manufacturers, through the supplemental NDA
process, to make minor modifications (i.e, those
not requiring manufacturers to sponsor new
clinical trials) in their products’ official indications for use. The extent to which marketed
drugs are used for unapproved indications is not
known. Manufacturers often have no economic
incentives to seek FDA approval for popular,
unofficial clinical uses of their products.
Second, the calculation of costs needed to
assess the cost effectiveness of a new drug would
require FDA to either: 1) use existing estimates
of costs of treating illnesses using alternative
forms of therapy, or 2) generate new data bases.
Problems encountered in using existing data
bases are discussed in a background paper of
this assessment .22 At present, FDA has no intramural capability for generating new data regarding the treatment costs for selected diseases.
Conceivably, FDA could ask or require drug
and medical device manufacturers’ to submit
estimates of the costs of treatments provided
during the clinical testing of a new drug. Meaningful estimates of treatment costs might be difficult to calculate, however, because the prices
charged for medical care are dynamic and can
vary substantially among geographical regions.
At the time FDA approves a drug or medical
device, it does not know—nor can it influence—
how much a manufacturer will charge for its
newly approved product. FDA could ask a manufacturer to estimate a product’s selling price.
Apparently, however, this task would be difficult for manufacturers to accomplish at the time
of market approval. Further, the price of a drug
or medical device will likely change over time,
and cost-effectiveness information calculated at
the time of market approval would likely
Third, the effect on competition of using cost
effectiveness as a criterion for market approval
of drugs and medical devices is unknown. Potential implications vary substantially depending on how FDA might use the cost-effectiveness
If FDA assumed the responsibility for conducting CEAS, theoretically manufacturers
would not be burdened with the expense of developing the capability to conduct such analyses. Most research-based companies, however,
would likely establish their own capabilities; at
least two pharmaceutical firms currently perform CEAS on some of their products. One company, Merck Sharp and Dohme, developed a
computerized cost-benefit model to illustrate the
costs and benefits of its pneumococcal vaccine
(464). Another pharmaceutical manufacturer,
SmithKline and French, has extensively studied
the costs and benefits of the use of one of its
drugs, cimetidine, in different populations
(529). If FDA required each sponsor of every
Ch. 8—Market Approval for Drugs and Medical Devices ● 97
new drug or medical device to conduct CEA as a
part of the premarketing approval processes for
drugs and devices, then the use of cost effectiveness would add to the financial and regulatory hurdles of new product introduction. Conceivably, that situation could reduce manufacturers’ willingness to bring a new product to
market and could reduce competition.
If FDA required only the original manufacturer of a new drug or medical device (i.e., one
not previously marketed in the United States) to
analyze the cost effectiveness of its product,
then the burden of conducting such analyses
would fall primarily on the leading innovating
If FDA used cost effectiveness as a criterion
for the marketing of only new therapeutic entities, and not for either generic or “me too” products, 23 and if manufacturers perceived such a
task as too expensive, then manufacturers might
attempt to break into existing markets for multiple source products rather than to develop new
drugs and create new markets.
If FDA applied cost-effectiveness criteria to
“me too” type products, then a manufacturer
might lower the introductory price of its new
product in order to make it compare favorably
to already-approved products in the same therapeutic category. Once a product was marketed,
however, its manufacturer could alter its price,
and FDA has no authority to control the price of
approved drugs and medical devices. Excessive
increases in postapproval prices would likely be
limited by competition in the marketplace for
similar products.
Fourth, the use of CEA to evaluate new drugs
and medical devices would require extensive resources, substantial time, and creative application of existing data. FDA would likely have to
compare the relative safety, efficacy, and cost of
an applicant product with those same characteristics of marketed products. In order to perform
such comparisons, FDA would have to do the
quantitatively assess the safety (risks) and
efficacy (benefits) of each marketed drug
and medical device;
establish standards for clinical efficacy,
safety, and cost of each available form of
treatment—and possibly prevention—in
numerous disease states (note: such standards could be incorporated into a monograph system such as those used for OTC
drug products and antibiotics); and
calculate cost-effectiveness information for
the use of each drug and medical device in
specialized populations, such as the elderly
and persons with specific medical problems.
Although it is debatable whether FDA has
statutory authority to evaluate new products
relative to currently marketed products, a provision in the Senate-passed version of the Drug
Regulation Reform Act of 1979 would allow
FDA to consider formally and explicitly the
“benefits and risks of available therapies” when
evaluating applicant drugs in the NDA review
Fifth, the cost effectiveness of different products would vary substantially depending on the
forms of treatment being compared. A comparison between a new drug and a surgical procedure in the treatment of a selected medical problem, for example, could yield large differences
in cost-effectiveness ratios. Such a comparison
might be useful to FDA in its evaluation of the
drug. The information yielded could be helpful
in the evaluation of the new product, especially
if the product were a new chemical entity that
represented a therapeutic breakthrough. Ultimate treatment for uncontrollable malignant
hypertension in a young person, for example,
has been the removal of one or both kidneys
(nephrectomy). Until recently, many cases of
this disease have not been controllable through
the use of drugs. A potent new antihypertensive
drug, minoxidil, has been shown to effectively
lower very high blood pressure, and in some
studies, its use reduced the need for nephrectomies (482). There is no other product on the
market that resembles minoxidil either chemically or therapeutically. When compared to sur-
gical removal of the kidneys, minoxidil would
likely be quite cost effective for some patients.
Comparisons between two drugs, however,
especially two drugs with similar therapeutic effects, would likely yield only small differences
in cost-effectiveness ratios. Small differences
would be of little value. Thiazide diuretics, for
example, represent another form of antihypertensive drug therapy. There are approximately
30 different single-entity thiazide diuretic products on the market, and each one produces very
similar therapeutic and adverse effects. In costeffectiveness comparisons among these 30 products, the differences in cost-effectiveness ratios
would likely be very small; hence such compari-
sons would likely be of little distinguishing value in FDA’s NDA review process for a new thiazide diuretic.
Sixth, the cost effectiveness of a drug or a
medical device might be influenced more by the
conditions of its use than by its demonstrated efficiency in premarketing clinical trials. Factors
such as dosage regimen, route of administration, and palatability (taste) all could influence
a product’s acceptance by patients and hence affect its cost effectiveness. Even if FDA were able
to accurately assess the cost effectiveness of a
drug or medical device at the time of market approval, such an assessment might not accurately
predict the efficiency of the product in general
Seventh, if FDA used cost effectiveness as a
criterion in its drug and medical device market
approval processes, the Agency would have to
rely on speculative and uncertain data for two
important valuations. First, FDA measures the
safety and efficacy of drugs and medical devices
in terms of changes in physiologic functions, not
in terms of changes in a person’s overall health
status. The efficacy of an antihypertensive drug,
for example, is evaluated on the degree to which
the drug reduces a person’s blood pressure. FDA
does not quantify the effect a drop in blood
pressure would likely produce in the health
status of a hypertensive patient. Second, FDA
analysts would face the challenge of assessing
the economic value of physiological changes—
and if possible of those in health status—among
persons using the medical device or drug.
Eighth, by using cost effectiveness as a criterion in the market approval process for drugs
and devices, FDA might be extending its role
beyond the scope of responsibilities Congress
intended the Agency to have. FDA’s primary
purpose is to protect the public from unsafe and
ineffective drugs and medical devices and from
unsafe cosmetics, foods, and food additives.
Thus far, Congress has not asked FDA to use
economic criteria to regulate the choice of safe
and effective products during the market approval process. Choices based on economic criteria are left to be made by consumers, health
care practitioners, hospital administrators, and
private and public health insurance carriers
after a product has been marketed.